STAT 330: 95-3

Assignment 7 (SAS Asst 1) Comments on Answers

1. Pregnancy Question: This is a clearly paired data problem so that the unpaired test is not valid. You use the data step to create a variable equal to the difference between the two pregnancy glucose levels. Then proc means provides means, SDs, standard errors and a t-test for the hypothesis that there is no difference in mean glucose levels. The P-value is just a bit below 0.05 (2 sided) so the conclusion is that there is a significant difference. The evidence is not too strong, however. The probability plots are satisfactory.

2. This is a two sample problem. In the file on disk there are, inexplicably, only four females. The null hypotheses of equal means and of equal variances are both accepted. The normal probability plots (you get two if you specify by sex in proc univariate) show no problem for men. For women with the truncated data set there is a bend in the plot by little conclusion can be drawn for only 4 data points.

3. The two sample t-test rejects the hypothesis of equal means whether you assume the variances are equal or not. The assumption of equal variances is accepted by any test and the normal probability plots are OK.

• About this document ...